THE FORMATION OF MELANINS. 355 



This formula for antialbumid is based upon an analysis of 

 antialbumid prepared by Kiihne and Chittenden * from serum- 

 albumin, while the formula for the melanin is based upon the 

 analysis of a product (melanoidic acid) prepared by Schmiede- 

 berg f by boiling serum-albumin with 25 per cent hydrochloric 

 acid for twelve hours. Schmiedeberg likewise prepared a 

 melanoidic acid from the fibrinoses contained in Witte's 

 " pepton " by heating these proteoses with phosphoric acid for 

 two months. Ascribing to the proteoses present in Witte's 

 "pepton " a formula with 102 atoms of carbon, he formulates 

 the production of the melanin as follows : 



C 102 H 150 K 3 oS0 81 + 18H 2 = C 102 H 98 Ni 2 S0 82 + 18NH, + 17H 2 0. 

 Fibrinose. Melanin. 



It is to be observed that these two artificial melanins or 

 melanoidic acids differ from each other by nearly 6 per cent 

 of carbon and 3 per cent of nitrogen, and that sulphur was 

 determined in only one of the products, the amount found 

 being 0.96 per cent. While the data may be considered as 

 perhaps hardly sufficient to justify such definite expressions 

 as the above formulae, the general trend of the argument is 

 exceedingly interesting and important. Any one who has 

 worked much with the proteids is aware how readily these 

 substances yield dark colored solutions on boiling with dilute 

 mineral acids ; a reaction not due to oxidation alone since it 

 takes place readily even hi the presence of reducing agents. 

 It is thus evident that melanin-like substances may be formed 

 from proteids by simple hydrolytic cleavage, and consequently 

 the question at once arises how far the nature of the mother 

 substance modifies the character of the resultant pigment. 

 Further, is one proteid better adapted for the artificial pro- 

 duction of a melanin than another proteid ? How closely do 

 the artificial melanins resemble in composition and general 

 characters the natural pigments of this class? And lastly, 

 how far can reactions of this kind be accepted as indicating 



* Kiihne and Chittenden, Zeitschrift fur Biologic, 1883, xix, p. 176. 

 t Schmiedeberg, loc. cit., p. 66. 



